Graduate School of Medicine, University of Tokyo, Tokyo, Japan.
Economics Faculty, Seijo University, Tokyo, Japan.
Front Public Health. 2022 Jul 12;10:913875. doi: 10.3389/fpubh.2022.913875. eCollection 2022.
To achieve the Sustainable Development Goals, strengthening investments in health service inputs has been widely emphasized, but less attention has been paid to tackling variation in the technical efficiency of services. In this study, we estimated the technical efficiency of local public health programs for the prevention of unintentional childhood injury and explored its contribution to national trend changes and regional health disparities in Japan. Efficiency scores were estimated based on the Cobb-Douglas and translog production functions using a true fixed effects model in a stochastic frontier analysis to account for unobserved time-invariant heterogeneity across prefectures. Using public data sources, we compiled panel data from 2001 to 2017 for all 47 prefectures in Japan. We treated disability-adjusted life years (DALYs) as the output, coverage rates of public health programs as inputs, and caregivers' capacity and environmental factors as constraints. To investigate the contribution of efficiency to trend changes and disparities in output, we calculated the predicted DALYs with several measures of inefficiency scores (2001 average, yearly average, and prefecture-year-specific estimates). In the translog model, mean efficiency increased from 0.62 in 2001 to 0.85 in 2017. The efficiency gaps among prefectures narrowed until 2007 and then remained constant until 2017. Holding inefficiency score constant, inputs and constraints contributed to improvements in average DALYs and widened regional gaps. Improved efficiency over the years further contributed to improvements in average DALYs. Efficiency improvement in low-output regions and stagnated improvement in high-output regions offset the trend of widening regional health disparities. Similar results were obtained with the Cobb-Douglas model. Our results demonstrated that assessing the inputs, constraints, output, and technical efficiency of public health programs could provide policy leverage relevant to region-specific conditions and performance to achieve health promotion and equity.
为实现可持续发展目标，人们广泛强调加强对卫生服务投入的投资，但对服务技术效率的变化关注较少。本研究旨在估计日本预防儿童意外伤害的地方公共卫生计划的技术效率，并探讨其对全国趋势变化和区域卫生差异的贡献。利用真固定效应模型在随机前沿分析中基于柯布-道格拉斯和超越对数生产函数来估计效率得分，以考虑到各都道府县之间不可观测的时间不变异质性。使用公共数据源，我们为日本所有 47 个都道府县编制了 2001 年至 2017 年的面板数据。我们将残疾调整生命年（DALY）作为产出，公共卫生计划的覆盖率作为投入，以及照顾者的能力和环境因素作为约束。为了研究效率对产出趋势变化和差异的贡献，我们使用几种无效分数（2001 年平均值、年平均值和都道府县-年份特定估计值）计算了预测的 DALY。在超越对数模型中，平均效率从 2001 年的 0.62 提高到 2017 年的 0.85。都道府县之间的效率差距在 2007 年前缩小，之后直到 2017 年保持不变。在保持无效分数不变的情况下，投入和约束对平均 DALY 的改善做出了贡献，并扩大了区域差距。多年来的效率提高进一步促进了平均 DALY 的改善。低产出地区的效率提高和高产出地区的停滞不前的改善抵消了区域卫生差异扩大的趋势。柯布-道格拉斯模型也得到了类似的结果。我们的研究结果表明，评估公共卫生计划的投入、约束、产出和技术效率可以为特定地区的情况和绩效提供相关的政策杠杆，以实现健康促进和公平。
